Support for LARGE or LBA drives? I've tried to find this out myself and only came up with some hacking guides for the Award bios.

Support for LARGE or LBA drives?

No, that came on the later 4.50 releases. I'm not sure on the exact versions, but I think the feature was introduced sometime in July or August '94, and they finally got it to actually work properly in the early '95 versions.

"Large" disks and LBA were already available on 4.50PG. In the past, I recall reading that some people wanted a MB with 4.51PG, but don't recall why.

Maybe 4.51PG is user-hackable to allow for up to 128 GB hard drives?

Think it would be feasible to compare the 2 binaries with a program like modbin or something?
For all we know the updated BIOS file only corrected a stupid spelling error 🤣!

Maybe, though I don't have any experience with modbin and BIOS HEX word modifications. This may be more of a whole evenings' chore that I'm not really up to.

If I had to speculate out on a limb, 4.51PG may allow for 32 GB hard drives whereas the 4.50PG allows for up to 8 GB IDE, PIO-4 hard drives.

My reasoning for this is not very solid, but I have had old 4.51PG BIOS's on Pentium boards which generally only support 32 GB hard drives. Some time, I'll test out this theory.

The most glaring difference I can remember of is the boot selection - usually v4.50 gives only A,C and C,A as boot priority. No Boot from CD option in BIOS in v4.50 iirc. I maybe wrong, though.

later 4.51 BIOSes support 128GB... the earlier ones theoretically did but it was bugged and can't work above 32GB... this was fixed later on... I have a hacked shuttle hot-569 bios that fixed the LBA problem and also improved K6-III functionality
I remember Award had a change log of their BIOSes on their website based on release dates... I don't know if it still exists ever since they got bought out by Phoenix

The latest 2002 BIOS from the Gigabyte GA-6BXC 440BX mainboard also says AWARD v4.51PG.

My 486 has an Award 4.51PG, and it only supports 8 GB IDE drives. It also can't boot from CD.

5u3, which 486 motherboard would that be?

More speculation: maybe 4.51PG supports a wider variety of PCI cards? I can't seem to find much information on AWARD revision history online.

Asus PVI-486SP3. Some of the limitations probably come from the onboard IDE controller (CMD640).

5u3 wrote:

Asus PVI-486SP3. Some of the limitations probably come from the onboard IDE controller (CMD640).

You might want to consider a different motherboard.

"CMD640, the CMD Technology Inc product 0640, is an IDE interface chip which has a data corruption bug."

Unless you can handle the possibility of powering up to a corrupted hard drive one day.

sliderider wrote:

You might want to consider a different motherboard.

"CMD640, the CMD Technology Inc product 0640, is an IDE interface chip which has a data corruption bug."

Unless you can handle the possibility of powering up to a corrupted hard drive one day.

Just wanted to point out that you can't conclude the BIOS features from the version number, because it also depends on the hardware.

I know enough about the CMD640 bug to avoid it 😉

Btw, has anybody ever seen a 486 board with support for more than 8 GB?

@5u3
That is an attractive motherboard, with features similar to that of this here Biostar MB8433-UUD v3.0 w/4.51PG, except yours has the SiS chipset.

I would be very curious to know:

1) With what Cyrix 5x86 enhancement features you have tried it with and with which CPU?

2) Does the PS/2 port work?

3) Can you enable Write-thru L2 cache when 512 KB cache is installed.

4) Does it run at 2-1-1-1 and 0WS/0WS well?

5) Any stability quirks that you've noticed?

6) Does it have a 1:1, 1:1/2, 1:2/3, FSB-to-PCI BIOS option?

7) Does it work with EDO RAM?

8) Since it has a SiS chipset, I assume it works with some graphics cards more modern than a Matrox G200. What is the most modern graphics card that you've been able to run in here? Voodoo3, Voodoo2, nVidia Riva TNT2 64?

Thanks!

As for the other comment, I have never owened an IDE harddrive between 9 and 32 GB, so I cannot test. I have an 80 GB IDE hard drive that I can test, but I'm 90% sure that size won't detect correctly. I've always used SCSI hard drives for 486 machines. The CPU load on a 486 with an IDE interface becomes your bottleneck otherwise.

I get about 32,000 KB/s using Adaptec's SCSI Benchmark (EZ SCSI 5.01) from a PCI Ultra2-LVD HBA.

1. Sorry, no experience with Cyrix CPUs. I only have Intel and AMD chips.
2. The PS/2 connector works. It uses the same pinout as other Asus boards.
3. No idea, since I have only 256 KB. Once I tried to upgrade to 1024 KB, but failed (probably faulty cache chips).
4. Yes.
5. Nothing out of the ordinary. On earlier board revisions the 3.3V voltage converter is a bit weak.
6. FSB=PCI, no dividers available.
7. No.
8. Haven't tried anything more modern than a Virge DX. 😊

feipoa wrote:

As for the other comment, I have never owened an IDE harddrive between 9 and 32 GB, so I cannot test. I have an 80 GB IDE hard drive that I can test, but I'm 90% sure that size won't detect correctly.

You can get a large HDD such as your 80 GB IDE model to be detected correctly by an older controller by "capping" its size using a piece of software such as the free to download SeaTools for DOS. I've used this utility quite a lot, and so far I haven't had any problems with it. I use the CD-ROM based .ISO download file, but I see on that webpage that a FDD image is also available.

If I use an old mobo that does not allow for CD-ROM boot up, I switch mobos to one that can, and then run SeaTools to "cap" the HDD's size, and then switch back to the mobos that I want to test it with. I think SeaTools for DOS requires a mouse to be present, and one that works with Ctmouse.exe.

feipoa wrote:

I've always used SCSI hard drives for 486 machines. The CPU load on a 486 with an IDE interface becomes your bottleneck otherwise.

I get about 32,000 KB/s using Adaptec's SCSI Benchmark (EZ SCSI 5.01) from a PCI Ultra2-LVD HBA.

I've never ventured in to the world of SCSI. Pity really, because they sound better than IDE. Does this CPU load with IDE bottleneck diminish, once you move away from 486s, and in to Pentium class systems? Also, please can you tell me how you benchmarked your SCSI system to reach a speed of 32,000 KB/s? I would like to run a comparitive test on a 486 with IDE, to see how far behind it is. Thanks a lot.

retro games 100 wrote:

Does this CPU load with IDE bottleneck diminish, once you move away from 486s, and in to Pentium class systems?

CPU load is reduced once you move from the programmed I/O (PIO) modes to (U)DMA access. This feature started to show up during the Pentium era (Socket 5,7).

Thanks for the info. I've got an idea. If you are lucky, and your mobo's BIOS doesn't balk at the presence of one, how about using a PCI-based IDE controller card, that supports (U)DMA access? If there is a vague hope that this hardware combo might work, I would like to test this. I could try this out on a fairly modern 486, such as a UMC chipset-based mobo. Any idea what "speed" PCI controller card I should be looking for please? As a guess, how about something like ultra 66? Thanks a lot.

UDMA IDE is still slower than SCSI because of overhead... CPU utilization is still higher due to how it handles data
the SCSI advantages only diminished once SATA came out.... there's little gained when going from SATA to SAS

my ancient seagate baracuda Ultra wide SCSI drive in my defunct *nix box still gives a UDMA/100 PATA drive a run for its money